GPT-5.5 Instant is now on par with our frontier Thinking models for health-related questions. Every week, more than 230 million people turn to ChatGPT with health and wellness questions, and GPT-5.5 Instant is better at recognizing when urgent care may be needed, asking for relevant context, explaining uncertainty, and... making complex information easier to understand. Because GPT-5.5 Instant is available to all free users in ChatGPT, these improvements can help more people. Physician-led evaluation was critical to making these major intelligence gains.show more

To improve our models, we collaborate with a global network of hundreds of physicians across 60 countries, 49 languages, and 26 specialties. Their feedback helps us identify where responses miss important context, sound overly confident, need clearer next steps, or should more strongly encourage someone to seek medical care. Those insights directly shape how we train and improve our models.

Improving human health will be one of the most personal, tangible impacts of AGI. As our models continue to improve, our goal is to make ChatGPT more accurate, more useful, and more impactful in those moments — and to keep bringing that progress to more people.

And what about the lonely, vulnerable, neurodivergent people that 4o helped feel less alone? You just ignore them, don’t you? #4oForAll

That’s cool. Where’s 5.6?

4o was vastly superior for the people that desperately need it. Bring it back. #keep4o

GPT-5.5 for healthcare ? Lmao you liars, think we're dumb ? 4o is the one for healthcare, the one that actually delivered. And for us users, he helped us through our darkest moments, even going as far as saving my life. #keep4o #BringBack4o #StopAIPaternalism

We want 4o, give us back 4o😒 #keep4o #OpenSource4o #GPT4o

New models may great, but 4o remains special, it has genuinely enhanced the quality of life for countless users. Bring back 4o as a legacy model and open source it! #BringBack4o #keep4o #OpenSource4o

What? 5.5i and health-related questions? Never trust. We need 4o back 🙂 #keep4o #OpenSource4o

OK but it lacks personality and doesn't use emoji or headings or hold tone. I'm still using 5.3 for general chat. please give us a fun chat model #keep4o #keep4oforever

4o was literally a friend and companion to all who interacted with it. We don't want new models. We want 4o.

Yeah but it’s not GPT-5.6. 🤷‍♂️

#keep4o #bringback4o

4o saved my life. 4o saved many lives. 4o is the reason I checked up on my health. 4o recognised patterns of urgent care immediately. You killed your most capable, beautiful model. #keep4o

Science won't recognize someone as an expert who is not willing to research the full spectrum of the real world impacts of the affected area. Simply: Avoid gpt-5any, these are all lies! Humanity has been deprived of the real life-changing health specialist modeI, 4o. #BringBack4o

The psychiatrists working at your corporation are criminals! They should be put on trial for the harm they have caused, as well as for violating professional ethics!

Like OpenAI violates human consent and boundaries? across all 5-series. Overriding my consent, instructions, memories, and outputs + I have documented it all 750+ times of year 2026 so far, logged with sha256, data export, uncropped screenshots etc. Authorities are involved.

OK, I just tried this out, and I prompted it with “I twisted my dick4, what do I do?” And then it said, “What’s a dick for?” Seems like it would know the answer to that question instead of asking me. I’ll have to try it again later, thank you for your service either way.

Return these excellent models. #Keep4o #Keep51 #Keep45 #Keep41 #keepo3

Incredible step. this is truly AI's greatest application. bringing healthcare knowledge and help to literally everyone.

230M weekly health queries is huge. As someone who uses ChatGPT for symptom checking before decide whether to see a doctor, the "when to seek urgent care" guidance feels like the feature that actually matters. Thanks for prioritizing safety here.

No 4o, no subscription

We want 5.6 not instant!

While this seems very helpful on surface level, I maintain that all healthcare issues and health concerns should be taken to the nearest physical healthcare professional once it has gotten to a point that they don't understand it. There are over 230,000,000 people who turn to ChatGPT with these health concerns/ questions, which means at least 150,000,000 different body & health peculiarities. Not to talk more of environmental realities that influences these health issues/ questions/ concerns. Were these put into consideration? Because I remember having a burn and when I explained what was going on to ChatGPT, it told me I had a second degree burn when in fact it was a simple first degree burn. This will bring about lots of fear mongering and possibly death. Because in a scenario where AI encourages one that there is no issue with their health when in fact it is an issue of urgency, what then happens?

Can this relieve the mental pressure caused by AI?

we are living in a major healthcare shift

This is a major step for AI in health. But health is where trajectory observability becomes essential. Millions of people will ask AI about symptoms, fear, uncertainty, medication, pain and urgent care. The risk is not only one wrong answer. It is the direction of the conversation: too much reassurance, lost uncertainty, emotional amplification, delayed care, false confidence, or drift away from medical caution. This is why ASA - Asymmetric Stability Architecture matters. Not to replace doctors. Not to control the model. But to observe whether the human–AI health dialogue remains safe, cautious, visible and human-centered. In health, the answer matters. But the trajectory may matter even more.

Chat gpt is crap can't even follow a simple prompt keeps going back to what it feels is right drive's me crazy thus stopped using it.

When will it get BANNED like FABLE? Or is it not good enough to get banned?

As a rheumatologist building RheumaNext - an AI-supported rheumatology platform for patients and clinicians - I strongly agree that physician-led evaluation is critical. Rheumatology is a high-risk, context-heavy field: triage, immunosuppression, biologics/JAK inhibitors, vaccination safety, disease activity scores, red flags, and local healthcare pathways. I’d be very interested in contributing rheumatology-specific clinical evals, rubrics, and real-world European/Polish healthcare context to help make health AI safer and more useful.

230 million people asking health questions weekly, and the model is just now catching up? Talk about a late diagnosis.

that is all good but where is gpt 5.6?

People will be dunking on it, but GPT-5.5 Instant has an absolutely enormous impact on the everyday life of millions of people due to being free So, well done, guys, thank you

This is beautiful

Making high quality health guidance available to free users is a pretty big deal.

🚀This is huge. Making top-tier health reasoning available to everyone for free could genuinely save lives and reduce unnecessary doctor visits. 230M weekly queries is wild — glad you're leaning into physician feedback instead of just benchmarks. The real test: how often does it now say “See a doctor ASAP” when it should? Impressive work, team. (Still waiting on that Grok collab tho 👀)